Tech volunteering: market failure?

It seems many EAs are technical. Also that many charities could benefit from more techies. Sounds like a market for volunteers, right? Cf. this post from Jeff Kaufman. Now Jeff was talking (primarily) about full time roles, in research orgs. I’m interested in something a little different: volunteering.

Lots of corporations have volunteering time opportunities. I know from personal observation and from conversations with others (Software and Tech Effective Altruism in London) that it is not easy to find volunteering opportunities that other EAs would agree were effective. In total, in the West I’d guess the pool of such under utilised days would be on the order of hundreds of thousands of professional techies times weeks.

This suggests a new organisation could be of use: a clearing house for effective tech volunteering opportunities. Match charities that have unfilled tech needs with corporates who might appreciate a steady stream of high quality volunteering opportunities.

Now, the question is, “What does a quality volunteering opportunity look like?” Let’s make a list of characteristics that come to mind:

  • It doesn’t have to be ground breaking. In particular: it doesn’t have to be machine learning.

  • It does need to have impact.

The first point could be expressed as, “It could be mundane.” For instance, I once saved someone a whole lot of time by writing a script to extract information from several hundred survey response spreadsheets. Sometimes the multiplier of tech comes from finding the most boring thing your people would rather not do, but cannot avoid because it is important and making a tool which automates it.

I comment on the ubiquity of ML from an explicitly personal point of view. I think a lot of the time people seem pretty eager to use the newest, biggest hammer on any particular problem. Only it’s very easy for people outside the field to misunderstand both how much data is needed and what kind of results it can produce. It’s awesome where it is well suited, but not every problem is an ML-nail.

The second point says that there is no point doing something cool which will not push forward the mission of the charity. Writing an AI to understand spoken commands of people in niche languages will likely be pointless if the intended users do not have smartphones or internet connections. It also needs to be used by the org: this is the curse of the hackathon project—cool things which stand alone and unused.

I suspect that, done well, such opportunities would incorporate engineering sprints into a relay of volunteers. If five people have five days, they may not be able to get much done individually, but collectively they could. Add in a project manager /​ scrum mistress who could actually keep track of a number of different projects, and modularise volunteer efforts on them. That would go a long way towards getting the proverbial done. And nothing said volunteers couldn’t help write the project plan either.

These considerations would be important to overcome another barrier: persuading good people who maybe aren’t EAs already to take up the opportunities. If talented people could contract at a rate of several hundred pounds a day, the problems need to be of commensurate scope to capture their attention. I’m going to predict that the more is asked the bigger the response will be. ‘We need a mobile payment system that works in as many countries as possible!’ is inherently a lot more interesting than, ‘Please change the colour of the button on this page’. Similarly: ‘Please advise us how to capture more operational data so that we can make better decisions in the future’ is more interesting than, ‘Could you make a graph of these survey responses.’